ISSN 1671-3710
CN 11-4766/R
主办:中国科学院心理研究所
出版:科学出版社

• •    

受控语义加工的脑网络机制:语义控制网络与一般控制网络的分工协作

居萍, 周子航, 杨晓虹   

  1. 中国人民大学心理学系, 北京 100872 中国
    中国人民大学社会心理科学与工程研究中心, 北京 100872 中国
    江苏师范大学江苏省语言能力协同创新中心, 江苏 221100 中国
  • 收稿日期:2026-08-01 修回日期:2026-10-13 接受日期:2026-10-23
  • 基金资助:
    中国人民大学科学研究基金(中央高校基本科研业务费专项资金资助)项目成果(23XNKJ01); 国家自然科学基金(31871108)

Brain Network Mechanisms Underlying Controlled Semantic Processing: Functional Differentiation and Coordination Between the Semantic Control and Domain-General Control Networks

JU Ping, ZHOU Zihang, YANG Xiaohong   

  1. Department of Psychology, Renmin University of China 100872, China
    Research Center for Social Psychological Science and Engineering, Renmin University of China 100872, China
    Jiangsu Collaborative Innovation Center for Language Ability, Jiangsu Normal University 221100, China
  • Received:2026-08-01 Revised:2026-10-13 Accepted:2026-10-23
  • Supported by:
    Fundamental Research Funds for the Central Universities, and the Research Funds of Renmin University of China(23XNKJ01); National Natural Science Foundation of China(31871108)

摘要: 受控语义加工是个体依据任务目标要求,对语义知识的检索、选择与抑制进行调节的过程。既有研究分别强调语义控制网络的特异性作用与一般控制网络的领域一般性作用,但二者如何分工协作尚缺乏系统解释。本文集中梳理任务激活、脑损伤患者、脑刺激与功能连接等多方面证据,比较两类网络的功能贡献、难度响应和信息编码方式。现有证据表明,一般控制网络主要维持目标与规则、配置注意与工作记忆资源、监测冲突并组织反应;语义控制网络则依据语义内容和语境,对分布式语义表征实施定向检索、选择和竞争抑制。两类网络随任务要求变化动态协作。基于此,本文提出受控语义加工的多层级协同框架,该框架提出,受控语义加工并非依赖单一控制网络,而是由一般控制网络与语义控制网络构成的层级化控制系统,与语义表征系统之间的动态协同实现,从而为理解领域一般性控制与语义特异性调节如何共同支持灵活语义行为提供了新的理论框架。

关键词: 受控语义加工, 语义控制网络, 一般控制网络, 脑网络机制

Abstract: Controlled semantic processing refers to the goal-directed regulation of semantic knowledge retrieval, selection, and inhibition. Previous studies have emphasized semantic-specific and domain-general control mechanisms, yet the functional differentiation and coordination of the corresponding networks have yet to be systematically characterized. This review integrates evidence from task-based neuroimaging, lesion studies, brain stimulation, and functional connectivity to compare the two networks’ functional roles, sensitivity to task difficulty, and representational properties. The reviewed evidence indicates that the general control network primarily supports task-level regulation, including the maintenance of task goals and rules, attentional allocation, working-memory updating, conflict monitoring, and response selection. By contrast, the semantic control network supports the controlled retrieval of weak or non-dominant information, the selection of task-relevant meanings or features, and the suppression of dominant but contextually irrelevant representations. The two networks interact dynamically as task demands change. Building on this evidence, we propose a multilevel coordination framework in which controlled semantic processing is supported by dynamic interactions between a hierarchical control system—comprising the domain-general and semantic control networks—and distributed semantic representation systems. This framework provides an integrative account of how domain-general control and semantic-specific regulation interact to support flexible semantic behaviour.

Key words: controlled semantic processing, semantic control network, domain-general control network, brain network mechanisms